Lecture 24
4/8/04 12:42
Definition of Artificial Intelligence
• “Artificial Intelligence (AI) is the part of
computer science concerned with designing
intelligent computer systems,
• that is, systems that exhibit the
characteristic we associate with intelligence
in human behavior —
• understanding language, learning,
reasoning, solving problems, and so on.”
Lecture 24
Artificial Intelligence
(S&G, §§13.1
–13.3)
§§13.1–
— Handbook of Artif. Intell., vol. I, p. 3
4/8/04
CS 100 - Lecture 24
1
3
The Turing Test is widely accepted
But controversial
Is it too hard?
Is it too easy?
If a machine could pass the Turing Test,
would we be justified in concluding it had a
mind?
• What do you think?
CS 100
CS 100 - Lecture 24
4
• In 1950 Turing wrote, “I believe that in
about fifty years’ time it will be possible to
programme computers, with a storage
capacity of about 109, to make them play
the imitation game so well that an average
interrogator will not have more than 70
percent chance of making the right
identification after five minutes of
questioning.”
• Hasn’t happened …
•
•
•
•
•
CS 100 - Lecture 24
4/8/04
Turing’s Optimism
Controversy
4/8/04
2
• Two humans and a computer play the
“imitation game”
• Interrogator communicates with other two
through an interface that hides superficial
differences
• Computer tries to fool interrogator into
thinking it is the human subject
• The human tries to help the interrogator to
conclude that the other subject is the
computer
• Could a machine think?
• What do we mean by “thinking”?
• How could we tell if a machine were
thinking?
• One operational approach to these issues is
the Turing Test
CS 100 - Lecture 24
CS 100 - Lecture 24
The Turing Test
“Thinking Machines”
4/8/04
4/8/04
5
4/8/04
CS 100 - Lecture 24
6
1
Lecture 24
4/8/04 12:42
Turing’s Optimism (2)
Major Kinds of Tasks
• “I believe that at the end of the century the
use of words and general educated opinion
will have altered so much that one will be
able to speak of machines thinking without
expecting to be contradicted.”
• Also hasn’t happened
• Nevertheless, computers have been
programmed to do many “intelligent” things
and AI techniques are used in many
applications
• Computational
4/8/04
4/8/04
CS 100 - Lecture 24
7
– calculation (in the broad sense)
– computers do better than people
• Reasoning
– planning
– making decisions
– computers follow rules mechanically, people follow
them intuitively
• Recognition & action
– humans (& other animals) do very well
– contemporary computers do poorly
Foundations of Intelligence
• Natural language
– allow the interrelation of any facts or
hypotheses
– “if all M are P and S is an M, then S is P”
– “inference engine”
– statements
• Formal logic
– formulas
• Knowledge
• Graphical
– general: “all birds have wings”
– specific: “Shakespeare wrote Hamlet”
– “knowledge base”
CS 100 - Lecture 24
– networks
• Pictorial
– images
9
4/8/04
• Spot is a dog
• Spot is brown
• Every dog has four
legs
• Every dog has a tail
• Every dog is a
mammal
• Every mammal is
warm-blooded
• Spot is a brown dog, and, like any dog, has four
legs and a tail. Also, like any dog, Spot is a
mammal, which means Spot is warm-blooded.
• More formally:
4/8/04
CS 100
Spot is a dog.
Spot is brown.
Every dog has four legs.
Every dog has a tail.
etc.
CS 100 - Lecture 24
CS 100 - Lecture 24
Specific propositions
General propositions
11
10
Formal Language
(Symbolic Logic)
Natural Language Representation
–
–
–
–
–
8
Knowledge Representation
• General inferential procedures
4/8/04
CS 100 - Lecture 24
4/8/04
• dog(Spot)
• brown(Spot)
• (∀x)(dog(x) →
four-legged(x))
• (∀x)(dog(x) → tail(x))
• (∀x)(dog(x) →
mammal(x))
• (∀x)(mammal(x) →
warm-blooded(x))
CS 100 - Lecture 24
12
2
Lecture 24
4/8/04 12:42
Graphical Representation
(Semantic Net)
mammal
Pictorial Representation
(Image-based)
warmblooded
• Much information is
implicit in an image
• But can be extracted
when needed
• Humans have
prototype images for
each basic category
• Brains use a kind of
analog computing for
image manipulation
four-legs
Example
Inference
dog
tail
Spot
4/8/04
brown
CS 100 - Lecture 24
13
4/8/04
CS 100 - Lecture 24
14
Desirable Characteristics of a
Knowledge Representation System
• Adequacy
– Can it capture all the relevant knowledge?
• Efficiency
– How large is the representation, and how easily can
implicit knowledge be extracted?
• Extendability
– Can we add new concepts, properties, individuals, rules
of inference?
• Appropriateness
– Is it well suited to the intended knowledge domain?
4/8/04
CS 100
CS 100 - Lecture 24
15
3